Effects of a Ferulate-Derived Dihydrobenzofuran Neolignan on Angiogenesis,Steroidogenesis,and Redox Status in a Swine Cell Model

Abstract

In the ongoing search for new therapeutic compounds, lignans and neolignans, which are widely distributed in plants, deserve special attention because of their interactions with several biological targets. Searching for potential antiangiogenic agents related to natural lignans/neolignans, we were attracted by a previously studied synthetic dihydrobenzofuran neolignan. We synthesized the compound by means of an eco-friendly, enzyme-mediated biomimetic coupling of the methyl ester of ferulic acid, and the present study was aimed to deeply investigate its effect in angiogenesis bioassays validated in our laboratory. In addition, a previously well-defined granulosa cell model was employed to evaluate the effect of dihydrobenzofuran neolignan on cell viability, steroidogenesis, and redox status. Present data support the antiangiogenic effect of this neolignan. Moreover, we demonstrate that, at least at the highest concentrations tested, dihydrobenzofuran neolignan affects granulosa cell viability and steroidogenesis. In addition, the compound inhibits generation of free radicals and stimulates scavenger enzyme activities. The present data, which are a further deepening of the evaluation of the biological activities of the dihydrobenzofuran lignan in well-defined cell models, are of interest and worthy of special attention.

Introduction

In recent years, a renewed attention has been devoted to natural products as leading compounds in drug discovery. Many natural products found in plants are potential therapeutic agents to be used in a wide range of pathologies. ¹ Phytochemicals include compounds with various biological properties, which have possibly evolved to allow plants to cope with environmental challenges, including exposure to radiation and toxins and defense against pests and infectious agents. Among phytochemicals, phenolic compounds, and in particular those found in food plants, have received particular attention in past years, in particular as antioxidant and cancer chemopreventive agents. ² However, because their bioavailability is generally poor, this observation has stimulated the search for synthetic analogues, to be employed as potential antitumor therapeutic agents.^3,4 Within the large family of polyphenols, lignans and neolignans are a distinct group; in fact, their biogenesis, based on oxidative coupling of radical monomeric species, leads to an impressive structural variety paralleled by an array of promising biological properties. In addition, analogues of the natural lignans or neolignans can be prepared by simple metal- or enzyme-mediated biomimetic synthesis.^4,5 In the frame of our interest in natural-derived compounds with antiproliferative and antiangiogenic properties, we have recently obtained the unusual benzoxanthene lignan 1 by biomimetic coupling of caffeic acid phenethyl ester (CAPE), a honeybee propolis constituent. ⁶ This compound was able to interact with DNA through intercalation and minor-groove binding, and it displays antiproliferative activity against tumor cells ⁷ and antiangiogenic properties. ⁸ The evaluation of small molecules with antiangiogenic activity is of increasing importance, given that new vessel growth accompanies several pathological processes such as diabetic retinopathy, arthritis, and solid tumor growth. ⁹ More recently, in the ongoing search for potential antiangiogenic agents related to natural lignans/neolignans, we were attracted by the antiangiogenic properties of the synthetic dihydrobenzofuran neolignan 2 obtained by Apers et al. ¹⁰ from the methyl ester of ferulic acid (4) and related to 3,4-di-O-methylcedrusin (6), a natural product isolated from sangre de drago, the popular name of different South American Croton species. ¹¹ The structures of compounds 1 to 6 are reported in Figure 1A .

Figure 1.

(A) Structures of compounds 1 to 6. (B) Synthesis of dihydrobenzofuran neolignan compound 2.

Therefore, the present study was conducted to deeply assess the effect of compound 2 on the angiogenic process. A tridimensional bioassay setup in our lab ⁸ was used, and vascular endothelial growth factor (VEGF) production was evaluated in swine granulosa cells. This in vitro model was employed because these cells are actively involved in the regulation of follicular vascularization, a rare example of physiological angiogenesis.^12,13 Because of the easy cell recovery, this previously well-defined experimental model^12,14,15 was also used to study the effect of compound 2 on cell viability and steroidogenesis, the main features of granulosa cell function. Finally, because several positive effects of ferulic acid (5) are related to its antioxidant properties, ¹⁶ and taking into account a very recent study on the enhanced antioxidant activity of the ferulic acid dimer, ¹⁷ we sought to determine also the effect of compound 2 on granulosa cell redox status. ¹⁸ To this purpose, compound 2 was synthesized through an eco-friendly, enzyme-mediated biomimetic coupling of the methyl ferulate and employed in this study.

Materials and Methods

Chemistry: Synthesis of Compounds 4 and 2

Laccase from Trametes versicolor laccase (TvL) and all reagents used in this study were obtained from Sigma (Milan, Italy, and St. Louis, MO) unless otherwise specified; ferulic acid was purchased from Fluka (Milan, Italy).

Synthesis of methyl ferulate (4)

To a solution of ferulic acid (322 mg, 1.66 mmol) in MeOH (100 mL), H₂SO₄ conc. (2 mL) was added; the resulting mixture was stirred for 2 h under reflux. After cooling at room temperature, 200 mL of ethyl acetate was added to the mixture. The organic phase was washed with an NaHCO₃ solution and saturated brine, dried over anhydrous Na₂SO₄, and the solvent evaporated under vacuum to yield the methyl ferulate 4 (yield 95%). ¹H NMR data of the product are in perfect agreement with previous data reported. ¹⁹

Synthesis of compound 2

The methyl ferulate (4; 100 mg, 0.48 mmol) was dissolved in ethyl acetate (35 mL), whereas TvL (10 U, 25 mg) was dissolved in acetate buffer (15 mL, 0.1 M, pH = 4.7). The biphasic system was stirred at room temperature for 20 h and monitored by thin layer chromatography. The reaction was quenched by phase separation followed by ethyl acetate extraction of the water solution, and the dried organic solvent was evaporated; the crude residue was purified by flash chromatography using a gradient of petroleum ether:dichloromethane as the eluent to give 2 (24% yield). Mass spectrometry and ¹H NMR data are in agreement with previously reported data. ¹⁹

Evaluation of the Effects of Compound 2 on Angiogenesis

Angiogenesis bioassay

Endothelial cell culture

An immortalized porcine aortic endothelial cell (AOC) line ²⁰ was generously provided by José Yelamos (Hospital Universitario Virgen de la Arrixaca, El Palmar, Murcia, Spain). In all experiments, AOCs at the 19th passage were used and seeded in culture medium (CM) composed of M199 supplemented with sodium bicarbonate (2.2 mg/mL), bovine serum albumin (BSA; 0.1%), penicillin (100 IU/mL), streptomycin (100 µg/mL), amphotericin B (2.5 µg/mL), selenium (5 ng/mL), and transferrin (5 µg/mL).

Three-dimensional endothelial cell culture on a fibrin gel support

The microcarrier-based fibrin gel angiogenesis assay was performed as described by Basini et al. ²¹ Briefly, 12.5 mg of gelatin-coated cytodex-3 microcarriers in 1 mL phosphate-buffered saline (PBS) was incubated for 3 h to hydrate. After two washings in PBS and one in CM, the microcarriers were put in flasks containg 5 mL CM; AOCs (5 × 10⁵) were added and cultured for 24 h to let the endothelial cells coat the microcarriers. For the fibrin gel preparation, 40 µL microcarriers covered by AOCs were pipetted into six-well plates containing a solution of fibrinogen (1 mg/mL PBS, pH 7.6), added to 1250 IU thrombine (250 µL). Fibrin gels were allowed to polymerize for 30 min at 37 °C and then equilibrated for 60 min with 2 mL M199. After a change of the medium, AOCs were treated with VEGF (100 ng/mL; PeproTech EC Ltd, London, UK) in the presence or absence of 0.1, 1, 10, and 100 µM of compound 2. Plates were incubated at 37 °C under humidified atmosphere (5% CO₂). AOCs were cultured for 96 h, totally renewing the treatment after 48 h.

Quantification of AOC growth on the fibrin gel matrix

Endothelial cell proliferation in the fibrin gel matrix was evaluated by means of the public domain National Institutes of Health Program Scion Image Beta 4.02 (Scion Corporation, MA, http://rsb.info.nih.gov/nih-image/). Ten pictures were taken for each gel at 48 and 96 h; images were converted into gray scale, resized to 50% (Paintbrush Software, MS Office), and saved as 24-bit Bitmap format compatible with Scion. The modified images were then imported into the program, and measurements were made by drawing the perimeter of the area occupied by the AOCs expressed as number of pixels. To validate the measurement of the area covered by AOCs in fibrin gels as a reliable method to evaluate cell proliferation, fibrin gels were stained by the nuclear dye bis-benzimide (Hoechst 33258, 20 µg/mL in PBS for 60 min) and examined by the fluorescence microscope. ²² This procedure was performed 20 times; for each experiment, the number of nuclei was counted under fluorescence, and pictures of the aera covered by AOCs were taken to measure the surface covered in the fibrin gel. A strong correlation was observed between the area covered by AOCs and the number of nuclei found in the same area (r = 0.96).

VEGF production

Granulosa cell collection

Swine ovaries were collected at a local slaughterhouse from Large White cross-bred gilts, parity = 0. The stage of the cycle was unknown. Follicles were classified on a dimension-based fashion. ¹⁵ The ovaries were placed into cold PBS (4 °C) supplemented with penicillin (500 IU/mL), streptomycin (500 µg/mL), and amphotericin B (3.75 µg/mL); maintained in a freezer bag; and transported to the laboratory within 1 h. After a series of washings with PBS and ethanol (70%), granulosa cells were aseptically harvested by aspiration of large follicles (>5 mm) with a 26-gauge needle and released in medium containing heparin (50 IU/mL), centrifuged for pelleting and then treated with 0.9% prewarmed ammonium chloride at 37 °C for 1 min to remove red blood cells. Cell number and viability were estimated using a hemocytometer under a phase contrast microscope after vital staining with trypan blue (0.4%) of an aliquot of the cell suspension. Cells were seeded in CM M199 supplemented with sodium bicarbonate (2.2 mg/mL), BSA (0.1%), penicillin (100 IU/mL), streptomycin (100 µg/mL), amphotericin B (2.5 µg/mL), selenium (5 ng/mL), and transferrin (5 µg/mL). Once seeded, cells were incubated in the presence or absence of compound 2 (0.1, 1, 10, and 100 µM) and maintained for 48 h at 37 °C under humidified atmosphere (5% CO₂). The treatment was identical for all experiments performed in this study.

Granulosa cell VEGF production

Granulosa cells (10⁶/1 mL CM + 1% fetal calf serum) were seeded in 24-well plates and incubated for 48 h, as indicated above. VEGF in culture media was quantified by an enzyme-linked immunosorbent assay (ELISA; Quantikine, R&D System, Minneapolis, MN). This assay, developed for human VEGF detection, has been validated for pig VEGF. ²³ The assay sensitivity was 8.74 pg/mL, and the inter- and intra-assay coefficients of variation were always less than 7%. A Victor Reader (Perkin Elmer, Boston, MA) set to read at a wavelength of 450 nm emission was used to quantify the reaction product.

Evaluation of the Effects of Compound 2 on Granulosa Cell Function

Granulosa cell viability

Granulosa cells (2 × 10⁵/200 µL CM) were seeded in 96-well plates. Cell viability was assayed using a bioluminescent assay (ATP-lites; Packard Bioscience, Groningen, the Netherlands), which measures intracellular adenosine triphosphate (ATP) levels as an indicator of cell numbers. ATP is a cell viability marker because it is present in all metabolically active cells, and the concentration declines very rapidly when the cells undergo necrosis or apoptosis. The ATP lite-M assay system is based on the production of light caused by the reaction of ATP with added luciferase and D-luciferin. This is illustrated in the following reaction scheme: ATP + D-luciferin + O₂ → luciferase, Mg²⁺ → oxyluciferin + AMP + PPi + CO₂ + light. The emitted light is proportional to the ATP concentration. Briefly, 50 µL of mammalian cell lysis solution was added to 100 µL of cell suspension, and the plate was shaken for 5 min in an orbital shaker at 700 rpm to lyse the cells and stabilize ATP. Then, 50 µL of substrate solution was added to the wells, and the microplate was shaken for 5 min in an orbital shaker at 700 rpm. The plate was placed in the dark for 10 min, and the luminescence was measured in a luminometer Victor Reader (Perkin Elmer). The results were recorded in counts per second, and the percentage of cell viability was calculated with reference to the negative control (cells without compound 2) for presentation purposes. The viable cell number for correcting data has been tested by recording the absorbance related to a standard curve prepared by culturing in quintuplicate granulosa cells at different plating densities (from 10³ to 10⁵/200 µL). The curve was repeated in four different experiments. The relationship between cell number and absorbance was linear (r = 0.9). Cell number/well was estimated from the resulting linear regression equation. The assay detection limit was 10³ cell/well, and the coefficients of variation were less than 5%.

Granulosa cell steroid production

Granulosa cells (10⁴/200 µL CM) were seeded in 96-well plates and supplemented with androstenedione (28 ng/mL). Culture media were then collected, frozen, and stored at −20 °C until progesterone (P4) and estradiol 17β (E2) determination by validated radioimmunoassays. ²⁴ P4 assay sensitivity and ED₅₀ were 0.24 and 1 nmol/L, respectively; E2 assay sensitivity and ED₅₀ were 0.05 and 0.2 nmol/L. The intra- and interassay coefficients of variation were less than 12% for both assays.

Evaluation of the Effects of Compound 2 on Granulosa Cell Redox Status

Superoxide (O₂^–) assay

O₂^– production was evaluated by WST-1 (4-[3-(4-iodophenyl) – 2-(4-nitrophenyl)-2H – 5-tetrazolium]-1,3-benzene disulfonate) test (Roche, Mannheim, Germany). The assay is based on the cleavage of the water-soluble tetrazolium salt, WST-1, to a yellow-orange, water-soluble formazan. Evidence exists that tetrazolium salts can be used as a reliable measure of intracellular O₂^– production. ²⁵ A total of 10⁴ cells/200 µL CM were seeded in 96-well plates and incubated for 48 h. During the last 4 h of incubation, 20 µL of WST-1 was added to cells, and absorbance was then determined using the Victor Reader (Perkin Elmer) at a wavelength of 450 nm against 620 nm. The coefficients of variation were less than 3%.

Hydrogen peroxide (H₂O₂) assay

H₂O₂ production was measured by an Amplex Red Hydrogen Peroxide Assay Kit (Molecular Probes, PoortGebouw, the Netherlands); the Amplex Red reagent reacts with H₂O₂ to produce resorufin, an oxidation product. Briefly, 2 × 10⁵ cells/200 µL CM were seeded in 96-well plates and incubated for 48 h. After centrifugation for 10 min at 400 × g, the supernatants were discarded, and cells were lysed adding cold Triton 1% in TRIS HCl (100 µL/well) and incubating on ice for 30 min. Undiluted cell lysates were used to perform the test and read against a standard curve of H₂O₂ ranging from 0.195 to 12.5 µM. The absorbance was determined with the Victor Reader (Perkin Elmer) using a 540-nm filter. The coefficients of variation were less than 5%.

Scavenging enzyme assays

Granulosa cells (2 × 10⁵/200 µL CM) were seeded in 96-well plates and incubated for 48 h. After centrifugation for 10 min at 400 × g, the supernatants were discarded, and cells were lysed adding cold Triton 1% in TRIS HCl (100 µL/10⁵ cells) and incubating on ice for 30 min. Superoxide dismutase (SOD), catalase, and peroxidase activities were assessed in cell lysates as described below.

SOD activity was determined by a SOD Assay Kit (Dojindo Molecular Technologies, Kumamoto, Japan). Cell lysates were tested without dilution, and a standard curve of catalase ranging from 0.156 to 20 U/mL was prepared. The colorimetric assay was performed by measuring formazan produced by the reaction between tetrazolium salt (WST-1) and superoxide anion (O₂^–), produced by the reaction of an exogenous xantine oxidase. The remaining O₂^– is an indirect hint of the endogenous SOD activity. The absorbance was determined with Victor Reader (Perkin Elmer) reading at 450 nm against 620 nm. The coefficients of variation were less than 3%.

Catalase activity was measured by an Amplex Red Catalase Assay Kit (Molecular Probes) based on the formation of an oxidation product (resorufin) derived from the reaction between H₂O₂ given in excess and the Amplex Red reagent in the presence of horseradish peroxidase. Cell lysates were diluted 1:10 to perform the test and read against a standard curve of catalase ranging from 62.5 to 1000 mU/mL. The absorbance was determined with Victor Reader (Perkin Elmer) using a 540 nm filter. The coefficients of variation were less than 5%.

Peroxidase activity was measured by an Amplex Red Peroxidase Assay Kit (Molecular Probes) based on the formation of an oxidation product (resorufin) derived from the reaction between H₂O₂ given in excess and the Amplex Red reagent. Cell lysates were used undiluted to perform the test and read against a standard curve of peroxidase ranging from 0.039 to 5 mU/mL. The absorbance was determined with Victor Reader (Perkin Elmer) using a 540 nm filter. The coefficients of variation were less than 5%.

Statistical Analysis

The experiments were repeated at least five times (six replicates/treatment). Experimental data are presented as mean ± SEM; statistical differences between treatments were calculated with analysis of variance using Statgraphics package (STSC Inc., Rockville, MD). When significant differences were found, means were compared by Scheffè’s F test; p values less than 0.05 were considered statistically significant.

Results

Chemistry: Synthesis of Compounds 4 and 2

We have synthesized compound 2 through an eco-friendly, biomimetic methodology based on enzyme-mediated oxidative coupling. This methodology avoids the use of metal-based oxidative agents and allows an easy purification of the product. The methyl ferulate was employed as substrate for the oxidative coupling carried out in the presence of laccase from TvL, according to Figure 1B , and afforded as main product the dihydrobenzofuran neolignan compound 2. In agreement with previously reported enzymatic dimerizations, ⁵ the lignan was obtained as a racemic mixture of trans-substituted enantiomers, and consequently, the absolute configuration of compound 2 is not reported here.

Evaluation of the Effects of Compound 2 on Angiogenesis

Effect in angiogenesis bioassay

Compound 2, from 1 to 100 µµM, inhibited angiogenesis (p < 0.001), whereas the lowest concentration tested was ineffective after both 48 and 96 h of incubation.

In particular, 1 and 10 µM exerted a dose-dependent effect (p < 0.001), which was significantly stronger (p < 0.001) after the 96 h culture. On the contrary, the highest concentration tested appeared to be significantly (p < 0.001) more effective after the 48 h incubation ( Fig. 2 ).

Figure 2.

Effect of the treatment with compound 2 (0, 0.1, 1, 10, or 100 µM) for (A) 48 h or (B) 96 h on aortic endothelial cell (AOC) growth in the angiogenesis bioassay. The absence of common letter on the bars indicates a significant difference (p < 0.001). (C) Phase contrast micrographs showing AOC growth at 48 and 96 h in fibrin gel matrix.

Effect on granulosa cell VEGF production

Basal VEGF production from granulosa cells amounted to 904 ± 72 pg/mL (mean ± SEM). Compound 2, although ineffective at the lowest concentration tested, appeared to significantly inhibit (p < 0.001) VEGF production in the 1 to 100 µM concentration range; 1 and 10 µM exerted the same effect ( Fig. 3A ).

Figure 3.

Effect of the treatment with compound 2 (0, 0.1, 1, 10, or 100 µM) for 48 h on VEGF production (A) and adenosine triphosphate content in swine granulosa cell culture (B). The absence of a common letter on the bars indicates a significant difference (p < 0.001).

Evaluation of the Effects of Compound 2 on Granulosa Cell Function

Granulosa cell viability

Granulosa cell viability was negatively affected by the highest concentrations (10 and 100 µM) of compound 2 (p < 0.001); 0.1 and 1 µM were ineffective ( Fig. 3B ).

Granulosa cell steroid production

E2 and P4 basal production by granulosa cells was 5.3 ± 0.9 and 86.1 ± 11.2 ng/mL, respectively. Both steroids were significantly (p < 0.001) stimulated by 100 µM of compound 2, whereas the other concentrations tested were ineffective ( Fig. 4 ).

Figure 4.

Effect of the treatment with compound 2 (0, 0.1, 1, 10, or 100 µM) for 48 h on swine granulosa cell (A) E2 or (B) P4 production. The absence of a common letter on the bars indicates a significant difference (p < 0.001).

Evaluation of the Effects of Compound 2 on Granulosa Cell Redox Status

Superoxide (O₂^–) production

The effect induced by compound 2 on superoxide anion generation appeared biphasic: the two lowest doses displayed an inhibitory effect of the same extent (p < 0.001), whereas the highest doses, in particular 100 µM (p < 0.001), exerted a stimulatory action ( Fig. 5A ).

Figure 5.

Effect of the treatment with compound 2 (0, 0.1, 1, 10, or 100 µM) for 48 h on (A) swine granulosa cell O₂^–or (B) H₂O₂ production. The absence of a common letter on the bars indicates a significant difference (p < 0.001).

Hydrogen peroxide (H₂O₂) production

Compound 2 modified hydrogen peroxide production only at the highest concentration examined (100 µM; p < 0.001), which induced a pronounced increase; the other doses were ineffective ( Fig. 5B ).

Scavenging enzyme activities

The activity of all the enzymatic scavengers examined was significantly increased by 100 µM of compound 2 (p < 0.001); the compound resulted ineffective at the other concentrations tested ( Fig. 6 ).

Figure 6.

Effect of the treatment with compound 2 (0, 0.1, 1, 10, or 100 µM) for 48 h on swine granulosa cell (A) superoxide dismutase, (B) catalase, or (C) peroxidase activities. The absence of a common letter on the bars indicates a significant difference (p < 0.001).

Discussion

Angiogenesis is a feature of a limited number of physiological processes, whereas pathological angiogenesis occurs in diseases such as cancer, rheumatoid arthritis, endometriosis, and diabetic retinopathy. Notably, because of the critical dependence of solid cancers on neoangiogenesis for growth, progression, and metastasis, it is not surprising that, in the past two decades, worldwide academic and industrial research has been focused on looking for angiotherapeutic agents to fight cancer. However, clinical trials with antiangiogenic modalities targeting VEGF/VEGFR2 signaling have shown limited efficacy and occasional toxic side effects. Therefore, screening strategies for herbal phytochemicals based on other signaling pathways have been explored. ¹⁹ Among these molecules, ferulic acid (5) has been documented to exert stimulating as well as inhibiting effects on the angiogenic process,^26,27 but the reasons for these conflicting reports are not clarified at present. Compound 2, previously synthesized by Apers at al., ¹⁰ has been shown to retain the antiangiogenic effect, as evaluated in the chorioallantoic membrane assay. In addition, the ferulic acid dimer 3 has been recently reported as an antioxidant compound with enhanced properties with respect to compound 5. ¹⁷ In this study, we have synthesized compound 2 by employing an enzyme-mediated, eco-friendly, and biomimetic process, with the aim of a deeper evaluation of the antiangiogenic and antioxidant properties of this dimer. Our present data, which confirm the inhibition of angiogenesis by means of different bioassays such as tridimensional fibrin gel and VEGF production by granulosa cells, support and further extend findings by Apers et al. ¹⁰

Moreover, our present research was aimed to better investigate the effects of compound 2 using the previously well-defined swine granulosa cell model.^8,28 Our data demonstrate that the two highest concentrations negatively affect granulosa cell viability, and the highest concentration potently stimulates granulosa cell steroidogenesis. The molecular mechanisms responsible for these effects are unknown at present and deserve further investigation. As for the effect on steroidogenesis, because both P4 and E2 synthesis was increased, compound 2 could have stimulated a common initial step in steroidogenesis, perhaps related to cholesterol metabolism; evidence demonstrates an effect of compound 5 at this level. ²⁹

Interestingly, compound 2, which inhibits free radical generation and stimulates enzyme scavenger activities, maintains the well-known antioxidant effects shown by its parent compound, compound 5. ¹⁶

Taken together, the results of this study represent a further deepening in the evaluation of the biological activities of the dihydrobenzofuran lignan compound 2 by means of well-defined cell models. These effects appear to be of interest and worthy of special attention; future research could be extended to other dihydrobenzofuran neolignans, a promising group of nature-derived compounds with potential for the design of new multitarget drugs.